Herman is right - as long as refmac reads your generated DRG that is the dictionary it will use..

My way of doing this sort of thing:

Start with COOT with the DRG coordinates and ProDRG dictionary and try a bit of real space refinement.
Coot should move the coodinates a bit and you should see something happen.

Then read out those new coordinates and run refmac. Again the coords should move a bit, and at lest the B factors should change. If REFMAC dooesnt like your naming it will report it..

Again if COOT accepted the dictionary the first time I cant see why it wont read it again. are you sure you are finding the same cif file?

You will have to read your own DRG into coot or otherwise it will find something which is not appropriate for your problem..
Eleeanor

On 5 March 2018 at 13:36, <[log in to unmask]> wrote:

PS: If you have two different home-brewn ligands, you have to rename one of them (pdb and cif), otherwise the same dictionary will be applied to two different ligands. Also make sure your cif file is a dictionary and not just a coordinate file.

HS

 

Von: Schreuder, Herman /DE
Gesendet: Montag, 5. März 2018 14:31
An: 'Colin Levy'; [log in to unmask]
Betreff: AW: [ccp4bb] Small molecule not refined in coot.

 

Hi Colin and Michel,

 

In my experience, both refmac and coot will use the most recently read-in cif dictionary and there is no need to try to find an unique identifier for each new ligand one uses. The new dictionary overrides the old one. Finding a unique identifier for each new ligand would be a terrible and unnecessary hassle.

 

What usually goes wrong is that the name of the residue in the pdb file does not match the name of the residue in the cif file. And the same holds for the atom names. I would check with an editor that the resdue and atom names of the pdb file do match the residue and atom names in the cif dictionary.

 

Best,

Herman

 

Von: CCP4 bulletin board [mailto:[log in to unmask]] Im Auftrag von Colin Levy
Gesendet: Montag, 5. März 2018 13:38
An: [log in to unmask]
Betreff: [EXTERNAL] Re: [ccp4bb] Small molecule not refined in coot.

 

Hi Michel,

 

If your ligand is designated as DRG in your pdb then refinement programs will anticipate that it is:

 

Chemical Description

Name

5,6-DIHYDRO-BENZO[H]CINNOLIN-3-YLAMINE

Formula

C12 H11 N3

Formal charge

0

Molecular weight

197.236 g/mol

Component type

NON-POLYMER

 

 

If this is not the case then you need to choose a unique identifier, this should then allow your cif library to be utilised appropriately during refinement.

 

Colin

 

 

 

Dr. Colin W. Levy

MIB G016

Tel.  0161 275 5090

Mob.07786 197 554
[log in to unmask]

 

 Manchester Institute of Biotechnology | University of Manchester | 3.020 Garside Building | 131 Princess Street | Manchester | M1 7DN


 

 

 

 

On 5 Mar 2018, at 12:30, M T <[log in to unmask]> wrote:

 

Dear all,

I am actually dealing with a structure containing an unnatural ligand.

I generated the pdb file by drawing it on PRODRG server, I did a manual pre-fit in the map using coot and I manually merged the pdf file of my protein with the one of the ligand.

After that I did few cycles of refinement using Refmac5, with my mtz, my merged pdb file and the cif library of the ligand as input files.

First I saw that even if Refmac "completed succesfully" it didn't modify coordinates of the ligand, second I saw a warning about the "DRG" molecule in the log of Refmac (WARNING: duplicated name of monomer DRG Last entry will be used.), third in Coot I cannot use "Real Space Refine Zone" with the ligand and when I try to import the CIF dictionary produced by PRODRG (through the server or even the one generated through CCP4 ProDrg), it causes Coot crash (** (coot-bin:4467): WARNING **: Widget not found: cif_dictionary_file_selector_create_molecule_checkbutton /programs/i386-mac/ccp4/7.0/ccp4-7.0/bin/coot: line 288:  4467 Segmentation fault: 11  $coot_bin "$@").

Clearly I am not a skilled user of all these programs, so I certainly did a/some mistake/s, but if someone can give me tips to be able to refine my ligand... Should I rename my DRG molecule, should I verify particular things, should I use other ways to generate the pdb and the library of my ligand?

Thank you.

 

P.S.: I am running CCP4 and coot using SBGRID on a Mac.

Michel